Thirty percent of African Americans (AAs) with Mild Cognitive Impairment (MCI) have (DM), which increases risk for progression to dementia. Poorly controlled DM magnifies this risk, but whether improving glycemic control reduces the risk is uncertain. Previous studies have lacked rigor to resolve this uncertainty. The proposed randomized controlled trial (RCT) studies a high-risk population, which confers greater rigor. This single-site, double-blind, active-control, phase II RCT will compare the efficacy of DM-Specific Behavioral Activation (DM-BA) vs. Enhanced Usual Care (EUC) to prevent decline in verbal memory (primary outcome) over 2 years in 200 AAs over age 65 years with amnestic multiple-domain MCI and poorly controlled DM. DM- BA is a behavioral treatment for DM, as well as a secondary prevention strategy for dementia. In DM-BA, race- concordant community health workers will: 1) deliver in-home DM education tailored to persons with MCI; 2) use action plans to reinforce DM self-care; 3) facilitate telehealth visits with a DM nurse educator to guide DM care and address participants? health beliefs; and 4) increase primary care physicians? (PCP) awareness of participants? cognitive deficits and health beliefs to optimize treatment of DM. The control treatment, EUC, is usual medical care enhanced with DM self-care education. Both DM-BA and EUC deliver DM education and have the same number of in-home treatment visits (i.e., 6 visits over 6 months, and 5 booster visits over the next 18 months). EUC, however, does not include DM-BA?s behavioral approach to improve glycemic control, telehealth visits, or PCP communication. The treatment comparison will identify DM-BA?s specific efficacy over and above EUC. Participants will be recruited from primary care practices. We will administer the Hopkins Verbal Learning Test-Revised (HVLT-R) (to assess verbal memory; the primary outcome) and the Uniform Data Set neuropsychological battery (to assess executive function, processing speed, language, visuospatial function, and global cognition; all exploratory outcomes) at baseline and months 6, 12, 18, and 24. The primary efficacy analysis will compare trajectories in HVLT-R Total Recall scores over 2 years by treatment group. We will also explore whether APOE genotype moderates treatment effects, and whether Optical Coherence Tomography measures of retinal Vessel Area Density (a proxy for cerebral microvascular disease) and/or Retinal Nerve Fiber Layer thickness (a proxy for cerebral neurodegeneration) mediate treatment effects. This RCT is innovative because it will clarify whether improving glycemic control prevents cognitive decline in a high risk population and identify possible treatment mechanisms. This RCT is significant because it targets two major problems facing older AAs (i.e., poor glycemic control and dementia). AAs? high risk for this comorbidity reflects the impact of cultural factors (e.g., health beliefs) and requires culturally relevant treatment. We have the experience and expertise to test this treatment, and the opportunity to change how DM is treated to prevent cognitive decline in AAs with MCI and DM and meet the goals of the National Alzheimer?s Project Act.